基于改进INFO-Bi-LSTM模型的SO_(2)排放质量浓度预测

针对火电机组SO_(2)排放质量浓度的影响因素众多,难以准确预测的问题,提出一种改进向量加权平均(weighted mean of vectors,INFO)算法与双向长短期记忆(bi-directional long short term memory,Bi-LSTM)神经网络相结合的预测模型(改进INFO-Bi-LSTM模型)。采用Circle混沌映射和反向学习产生高质量初始化种群,引入自适应t分布提升INFO算法跳出局部最优解和全局搜索的能力。选取改进INFO-Bi-LSTM模型和多种预测模型对炉内外联合脱硫过程中4种典型工况下的SO_(2)排放质量浓度进行预测,将预测结果进行验证对比。结果表明:改进INFO算法的寻优能力得到提升,并且改进INFO-Bi-LSTM模型精度更高,更加适用于SO_(2)排放质量浓度的预测,可为变工况下的脱硫控制提供控制理论支撑。

基于Bi-LSTM与状态约束的心音分割算法

心音分割是进行准确心音分类的前提。针对心音分割,提出一种基于双向长短时记忆网络(Bi-LSTM)与状态约束的算法。该文通过网格法确定Bi-LSTM网络中的最佳参数,并训练出心音状态识别模型;统计Bi-LSTM预测的心音状态持续时间,并计算自相关参数;利用自相关参数和心音固有状态转移规则对预测的心音状态进行约束处理。使用五折交叉验证法在PhysioNet/CinC 2016数据集上进行测试,该算法与同类算法相比,整体性能更佳。

基于Bi-LSTM的浅层地下双孔洞探测技术

文章探究一种基于深度学习的浅层地下孔洞探测技术,以应对地下孔洞给桩基施工安全所造成的严重威胁。基于浅层地震反射波法的原理,采用基础施工过程中的桩锤激震作为激励源,通过在探测区域地表上布置少量加速度传感器采集孔洞反射信号,并将反射信号作为深度学习的输入,以输出孔洞信息,建立一种新型的智能孔洞探测方法。结果表明,双向长短期记忆神经网络(bidirectional long short-term memory neural network,Bi-LSTM)的预测模型对于地下双孔洞的工况具有较高的识别准确率,在容许误差为2 m的情况下,孔洞位置和直径的预测准确率可达95.3%。该研究验证了基于深度学习的多孔洞探测技术的可行性,有望为施工前期土层地质状况的评估提供技术保障。

基于贝叶斯自优化Bi-LSTM组合网络的高速铁路轨道-桥梁系统震后响应预测方法

中国高速铁路(HSR)规划建设逐渐向地震易发地区延伸,亟需一种及时、准确的灾后地震响应快速预测方法,以实现高速铁路系统运输生命线安全的快速评估。本文提出了一种基于贝叶斯自优化双向长短期记忆(Bi-LSTM)网络的快速预测方法,以经过实验验证的高速铁路轨道-桥梁系统有限元模型地震动响应计算数据为样本,将预测地震响应和有限元计算结果进行比较,验证所提方法的精度和鲁棒性,表明该方法在预测高速铁路桥梁结构的非线性地震反应方面是有效的,且高速铁路轨道-桥梁系统的不同预测位置对预测精度的影响不明显;此外,为了降低神经网络训练数据量需求,提出了一种基于离散小波分解的分层聚类算法,结果表明,基于小波分解的分层聚类方法在保证预测精度的同时,有效地减少了训练地震集的输入数量。

基于Bi-LSTM模型的恶意JavaScript代码检测方法

传统的静态检测恶意JavaScript代码方法十分依赖于已有的恶意代码特征,无法有效提取混淆恶意代码特征,导致检测混淆恶意JavaScript代码的精确率低。针对该问题提出基于双向长短期记忆网络(Bidirectional Long Short-term Memory, Bi-LSTM)的恶意代码检测模型。通过抽象语法树将JavaScript代码转化为句法单元序列,通过Doc2Vec算法将句法单元序列用分布式向量表示,将句向量矩阵送入Bi-LSTM模型进行检测。实验结果表明,该方法对于混淆恶意JavaScript代码具有良好的检测效果且检测效率高,准确率为97.03%,召回率为97.10%。

基于ICEEMDAN模糊熵与Bi-LSTM的工业设备健康状态预测

工业设备健康状态关系着工业生产能否正常进行,为此提出一种基于改进自适应噪声完备经验模态分解(ICEEMDAN)和双向长短期记忆网络(Bi-LSTM)的工业设备健康状态预测方法。ICEEMDAN用于将原始音频信号进行分解得到若干个固有模态函数(IMF)分量,通过计算相关系数选取最佳分量组进行信号重构,然后计算重构IMF分量的模糊熵值构造特征向量集合,最后再输入到Bi-LSTM网络进行模型训练和预测。实验结果表明:相较于其他模型,基于ICEEMDAN模糊熵和Bi-LSTM的工业设备健康状态预测方法,能够有效提取音频信号特征,并准确进行健康状态预测。

Pluggable multitask diffractive neural networks based on cascaded metasurfaces

Optical neural networks have significant advantages in terms of power consumption,parallelism,and high computing speed,which has intrigued extensive attention in both academic and engineering communities.It has been considered as one of the powerful tools in promoting the fields of imaging processing and object recognition.However,the existing optical system architecture cannot be reconstructed to the realization of multi-functional artificial intelligence systems simultaneously.To push the development of this issue,we propose the pluggable diffractive neural networks(P-DNN),a general paradigm resorting to the cascaded metasurfaces,which can be applied to recognize various tasks by switching internal plug-ins.As the proof-of-principle,the recognition functions of six types of handwritten digits and six types of fashions are numerical simulated and experimental demonstrated at near-infrared regimes.Encouragingly,the proposed paradigm not only improves the flexibility of the optical neural networks but paves the new route for achieving high-speed,low-power and versatile artificial intelligence systems.

Screening biomarkers for spinal cord injury using weighted gene co-expression network analysis and machine learning

Immune changes and inflammatory responses have been identified as central events in the pathological process of spinal co rd injury.They can greatly affect nerve regeneration and functional recovery.However,there is still limited understanding of the peripheral immune inflammato ry response in spinal cord inju ry.In this study.we obtained microRNA expression profiles from the peripheral blood of patients with spinal co rd injury using high-throughput sequencing.We also obtained the mRNA expression profile of spinal cord injury patients from the Gene Expression Omnibus(GEO)database(GSE151371).We identified 54 differentially expressed microRNAs and 1656 diffe rentially expressed genes using bioinformatics approaches.Functional enrichment analysis revealed that various common immune and inflammation-related signaling pathways,such as neutrophil extracellular trap formation pathway,T cell receptor signaling pathway,and nuclear factor-κB signal pathway,we re abnormally activated or inhibited in spinal cord inju ry patient samples.We applied an integrated strategy that combines weighted gene co-expression network analysis,LASSO logistic regression,and SVM-RFE algorithm and identified three biomarke rs associated with spinal cord injury:ANO10,BST1,and ZFP36L2.We verified the expression levels and diagnostic perfo rmance of these three genes in the original training dataset and clinical samples through the receiver operating characteristic curve.Quantitative polymerase chain reaction results showed that ANO20 and BST1 mRNA levels were increased and ZFP36L2 mRNA was decreased in the peripheral blood of spinal cord injury patients.We also constructed a small RNA-mRNA interaction network using Cytoscape.Additionally,we evaluated the proportion of 22 types of immune cells in the peripheral blood of spinal co rd injury patients using the CIBERSORT tool.The proportions of naive B cells,plasma cells,monocytes,and neutrophils were increased while the proportions of memory B cells,CD8^(+)T cells,resting natural killer cells,resting dendritic cells,and eosinophils were markedly decreased in spinal cord injury patients increased compared with healthy subjects,and ANO10,BST1 and ZFP26L2we re closely related to the proportion of certain immune cell types.The findings from this study provide new directions for the development of treatment strategies related to immune inflammation in spinal co rd inju ry and suggest that ANO10,BST2,and ZFP36L2 are potential biomarkers for spinal cord injury.The study was registe red in the Chinese Clinical Trial Registry(registration No.ChiCTR2200066985,December 12,2022).

Social-ecological perspective on the suicidal behaviour factors of early adolescents in China:a network analysis

Background In early adolescence,youth are highly prone to suicidal behaviours.Identifying modifiable risk factors during this critical phase is a priority to inform effective suicide prevention strategies.Aims To explore the risk and protective factors of suicidal behaviours(ie,suicidal ideation,plans and attempts)in early adolescence in China using a social-ecological perspective.Methods Using data from the cross-sectional project‘Healthy and Risky Behaviours Among Middle School Students in Anhui Province,China',stratified random cluster sampling was used to select 5724 middle school students who had completed self-report questionnaires in November 2020.Network analysis was employed to examine the correlates of suicidal ideation,plans and attempts at four levels,namely individual(sex,academic performance,serious physical llness/disability,history of self-harm,depression,impulsivity,sleep problems,resilience),family(family economic status,relationship with mother,relationship with father,family violence,childhood abuse,parental mental illness),school(relationship with teachers,relationship with classmates,school-bullying victimisation and perpetration)and social(social support,satisfaction with society).Results In total,37.9%,19.0%and 5.5%of the students reported suicidal ideation,plans and attempts in the past 6 months,respectively.The estimated network revealed that suicidal ideation,plans and attempts were collectively associated with a history of self-harm,sleep problems,childhood abuse,school bullying and victimisation.Centrality analysis indicated that the most influential nodes in the network were history of self-harm and childhood abuse.Notably,the network also showed unique correlates of suicidal ideation(sex,weight=0.60;impulsivity,weight=0.24;family violence,weight=0.17;relationship with teachers,weight=-0.03;school-bullying perpetration,weight=0.22),suicidal plans(social support,weight=-0.15)and suicidal attempts(relationship with mother,weight=-0.10;parental mental llness,weight=0.61).Conclusions This study identified the correlates of suicidal ideation,plans and attempts,and provided practical implications for suicide prevention for young adolescents in China.Firstly,this study highlighted the importance of joint interventions across multiple departments.Secondly,the common risk factors of suicidal ideation,plans and attempts were elucidated.Thirdly,this study proposed target interventions to address the unique influencing factors of suicidal ideation,plans and attempts.

基于注意力机制与XBOA-Bi-LSTM的离心式压缩机故障预警方法

由于离心式压缩机存在着运行工况复杂、维修成本昂贵和长输管道工作环境恶劣的问题,为此,提出了一种基于注意力机制(AM)和蝴蝶算法优化双向长短期记忆神经网络(XBOA-Bi-LSTM)的离心式压缩机故障预警方法。首先,针对传统蝴蝶算法的收敛速度慢、转换概率单一和容易陷入局部最优等问题,通过引入无限折叠迭代混叠映射以丰富蝴蝶算法的初始种群;同时,提出了一种基于种群离散度与迭代次数的自适应惯性转换概率,以提高蝴蝶算法的寻优能力;然后,采用了灰色关联度分析法对测点数据进行了特征提取,结合注意力机制对输入序列进行了灰色关联度系数赋权;最后,建立了双向长短期记忆神经网络故障预警模型,采用仿真实验完成了对离心式压缩机的故障预警;以某天然气长输管道机组的离心式压缩机作为仿真对象,对该离心式压缩机故障预警方法的可行性进行了验证。研究结果表明:采用基于注意力机制与XBOA-Bi-LSTM的离心式压缩机故障预警方法时,在离心式压缩机故障发生前2 h~3 h内就发出预警信号,实现了对于离心式压缩机进气过滤器压差异常与支撑轴承工作异常的故障预警目的。

Image super‐resolution via dynamic network

Convolutional neural networks depend on deep network architectures to extract accurate information for image super‐resolution.However,obtained information of these con-volutional neural networks cannot completely express predicted high‐quality images for complex scenes.A dynamic network for image super‐resolution(DSRNet)is presented,which contains a residual enhancement block,wide enhancement block,feature refine-ment block and construction block.The residual enhancement block is composed of a residual enhanced architecture to facilitate hierarchical features for image super‐resolution.To enhance robustness of obtained super‐resolution model for complex scenes,a wide enhancement block achieves a dynamic architecture to learn more robust information to enhance applicability of an obtained super‐resolution model for varying scenes.To prevent interference of components in a wide enhancement block,a refine-ment block utilises a stacked architecture to accurately learn obtained features.Also,a residual learning operation is embedded in the refinement block to prevent long‐term dependency problem.Finally,a construction block is responsible for reconstructing high‐quality images.Designed heterogeneous architecture can not only facilitate richer structural information,but also be lightweight,which is suitable for mobile digital devices.Experimental results show that our method is more competitive in terms of performance,recovering time of image super‐resolution and complexity.The code of DSRNet can be obtained at https://github.com/hellloxiaotian/DSRNet.

Mapping Network-Coordinated Stacked Gated Recurrent Units for Turbulence Prediction

Accurately predicting fluid forces acting on the sur-face of a structure is crucial in engineering design.However,this task becomes particularly challenging in turbulent flow,due to the complex and irregular changes in the flow field.In this study,we propose a novel deep learning method,named mapping net-work-coordinated stacked gated recurrent units(MSU),for pre-dicting pressure on a circular cylinder from velocity data.Specifi-cally,our coordinated learning strategy is designed to extract the most critical velocity point for prediction,a process that has not been explored before.In our experiments,MSU extracts one point from a velocity field containing 121 points and utilizes this point to accurately predict 100 pressure points on the cylinder.This method significantly reduces the workload of data measure-ment in practical engineering applications.Our experimental results demonstrate that MSU predictions are highly similar to the real turbulent data in both spatio-temporal and individual aspects.Furthermore,the comparison results show that MSU predicts more precise results,even outperforming models that use all velocity field points.Compared with state-of-the-art methods,MSU has an average improvement of more than 45%in various indicators such as root mean square error(RMSE).Through comprehensive and authoritative physical verification,we estab-lished that MSU’s prediction results closely align with pressure field data obtained in real turbulence fields.This confirmation underscores the considerable potential of MSU for practical applications in real engineering scenarios.The code is available at https://github.com/zhangzm0128/MSU.

Applying an Improved Dung Beetle Optimizer Algorithm to Network Traffic Identification

Network traffic identification is critical for maintaining network security and further meeting various demands of network applications.However,network traffic data typically possesses high dimensionality and complexity,leading to practical problems in traffic identification data analytics.Since the original Dung Beetle Optimizer(DBO)algorithm,Grey Wolf Optimization(GWO)algorithm,Whale Optimization Algorithm(WOA),and Particle Swarm Optimization(PSO)algorithm have the shortcomings of slow convergence and easily fall into the local optimal solution,an Improved Dung Beetle Optimizer(IDBO)algorithm is proposed for network traffic identification.Firstly,the Sobol sequence is utilized to initialize the dung beetle population,laying the foundation for finding the global optimal solution.Next,an integration of levy flight and golden sine strategy is suggested to give dung beetles a greater probability of exploring unvisited areas,escaping from the local optimal solution,and converging more effectively towards a global optimal solution.Finally,an adaptive weight factor is utilized to enhance the search capabilities of the original DBO algorithm and accelerate convergence.With the improvements above,the proposed IDBO algorithm is then applied to traffic identification data analytics and feature selection,as so to find the optimal subset for K-Nearest Neighbor(KNN)classification.The simulation experiments use the CICIDS2017 dataset to verify the effectiveness of the proposed IDBO algorithm and compare it with the original DBO,GWO,WOA,and PSO algorithms.The experimental results show that,compared with other algorithms,the accuracy and recall are improved by 1.53%and 0.88%in binary classification,and the Distributed Denial of Service(DDoS)class identification is the most effective in multi-classification,with an improvement of 5.80%and 0.33%for accuracy and recall,respectively.Therefore,the proposed IDBO algorithm is effective in increasing the efficiency of traffic identification and solving the problem of the original DBO algorithm that converges slowly and falls into the local optimal solution when dealing with high-dimensional data analytics and feature selection for network traffic identification.

Computing Power Network:A Survey

With the rapid development of cloud computing,edge computing,and smart devices,computing power resources indicate a trend of ubiquitous deployment.The traditional network architecture cannot efficiently leverage these distributed computing power resources due to computing power island effect.To overcome these problems and improve network efficiency,a new network computing paradigm is proposed,i.e.,Computing Power Network(CPN).Computing power network can connect ubiquitous and heterogenous computing power resources through networking to realize computing power scheduling flexibly.In this survey,we make an exhaustive review on the state-of-the-art research efforts on computing power network.We first give an overview of computing power network,including definition,architecture,and advantages.Next,a comprehensive elaboration of issues on computing power modeling,information awareness and announcement,resource allocation,network forwarding,computing power transaction platform and resource orchestration platform is presented.The computing power network testbed is built and evaluated.The applications and use cases in computing power network are discussed.Then,the key enabling technologies for computing power network are introduced.Finally,open challenges and future research directions are presented as well.

IDS-INT:Intrusion detection system using transformer-based transfer learning for imbalanced network traffic

A network intrusion detection system is critical for cyber security against llegitimate attacks.In terms of feature perspectives,network traffic may include a variety of elements such as attack reference,attack type,a subcategory of attack,host information,malicious scripts,etc.In terms of network perspectives,network traffic may contain an imbalanced number of harmful attacks when compared to normal traffic.It is challenging to identify a specific attack due to complex features and data imbalance issues.To address these issues,this paper proposes an Intrusion Detection System using transformer-based transfer learning for Imbalanced Network Traffic(IDS-INT).IDS-INT uses transformer-based transfer learning to learn feature interactions in both network feature representation and imbalanced data.First,detailed information about each type of attack is gathered from network interaction descriptions,which include network nodes,attack type,reference,host information,etc.Second,the transformer-based transfer learning approach is developed to learn detailed feature representation using their semantic anchors.Third,the Synthetic Minority Oversampling Technique(SMOTE)is implemented to balance abnormal traffic and detect minority attacks.Fourth,the Convolution Neural Network(CNN)model is designed to extract deep features from the balanced network traffic.Finally,the hybrid approach of the CNN-Long Short-Term Memory(CNN-LSTM)model is developed to detect different types of attacks from the deep features.Detailed experiments are conducted to test the proposed approach using three standard datasets,i.e.,UNsWNB15,CIC-IDS2017,and NSL-KDD.An explainable AI approach is implemented to interpret the proposed method and develop a trustable model.

基于TDCSO优化CNN-Bi-LSTM网络的井底钻压预测方法

为了准确预测井底钻压,提高钻井效率、降低钻井成本,建立了融合双向长短期记忆网络(Bi-LSTM)和卷积神经网络(CNN)的混合模型。采用三角函数驱动的粒子群优化(TDCSO)方法对模型进行超参数优化,以提高预测钻压的精度;采用美国犹他州FORGE 58-32井和FORGE 58-62井的2个公开数据集对建立的模型进行验证,并采用平均绝对误差、均方根误差、决定系数和均方误差等指标进行模型性能评估。研究结果表明,所提出TDCSO-CNN-Bi-LSTM模型平均绝对误差、均方误差和均方根误差等3个关键性能指标较好,其中决定系数大于0.980,明显优于现有的LSTM、GRU、CNN-LSTM、CNN-Bi-LSTM等方法。研究表明,所提出的TDCSO-CNN-Bi-LSTM模型在井底钻压预测方面具有出色的准确性,能够实现实时监测,并与自动钻进系统集成,实现对钻压的精准控制,不仅提高了钻井效率,还降低了钻井成本,对未来的钻井作业具有重要的实际应用价值。

Activation Redistribution Based Hybrid Asymmetric Quantization Method of Neural Networks

The demand for adopting neural networks in resource-constrained embedded devices is continuously increasing.Quantization is one of the most promising solutions to reduce computational cost and memory storage on embedded devices.In order to reduce the complexity and overhead of deploying neural networks on Integeronly hardware,most current quantization methods use a symmetric quantization mapping strategy to quantize a floating-point neural network into an integer network.However,although symmetric quantization has the advantage of easier implementation,it is sub-optimal for cases where the range could be skewed and not symmetric.This often comes at the cost of lower accuracy.This paper proposed an activation redistribution-based hybrid asymmetric quantizationmethod for neural networks.The proposedmethod takes data distribution into consideration and can resolve the contradiction between the quantization accuracy and the ease of implementation,balance the trade-off between clipping range and quantization resolution,and thus improve the accuracy of the quantized neural network.The experimental results indicate that the accuracy of the proposed method is 2.02%and 5.52%higher than the traditional symmetric quantization method for classification and detection tasks,respectively.The proposed method paves the way for computationally intensive neural network models to be deployed on devices with limited computing resources.Codes will be available on https://github.com/ycjcy/Hybrid-Asymmetric-Quantization.

Insights into microbiota community dynamics and flavor development mechanism during golden pomfret(Trachinotus ovatus)fermentation based on single-molecule real-time sequencing and molecular networking analysis

Popular fermented golden pomfret(Trachinotus ovatus)is prepared via spontaneous fermentation;however,the mechanisms underlying the regulation of its flavor development remain unclear.This study shows the roles of the complex microbiota and the dynamic changes in microbial community and flavor compounds during fish fermentation.Single-molecule real-time sequencing and molecular networking analysis revealed the correlations among different microbial genera and the relationships between microbial taxa and volatile compounds.Mechanisms underlying flavor development were also elucidated via KEGG based functional annotations.Clostridium,Shewanella,and Staphylococcus were the dominant microbial genera.Forty-nine volatile compounds were detected in the fermented fish samples,with thirteen identified as characteristic volatile compounds(ROAV>1).Volatile profiles resulted from the interactions among the microorganisms and derived enzymes,with the main metabolic pathways being amino acid biosynthesis/metabolism,carbon metabolism,and glycolysis/gluconeogenesis.This study demonstrated the approaches for distinguishing key microbiota associated with volatile compounds and monitoring the industrial production of high-quality fermented fish products.

Five commonly used traditional Chinese medicine formulas in the treatment of ulcerative colitis:A network meta-analysis

BACKGROUND Currently,traditional Chinese medicine(TCM)formulas are commonly being used as adjunctive therapy for ulcerative colitis in China.Network meta-analysis,a quantitative and comprehensive analytical method,can systematically compare the effects of different adjunctive treatment options for ulcerative colitis,providing scientific evidence for clinical decision-making.AIM To evaluate the clinical efficacy and safety of commonly used TCM for the treatment of ulcerative colitis(UC)in clinical practice through a network metaanalysis.METHODS Clinical randomized controlled trials of these TCM formulas used for the adjuvant treatment of UC were searched from the establishment of the databases to July 1,2022.Studies that met the inclusion criteria were screened and evaluated for literature quality and risk of bias according to the Cochrane 5.1 standard.The methodological quality of the studies was assessed using ReviewManager(RevMan)5.4,and a funnel plot was constructed to test for publication bias.ADDIS 1.16 statistical software was used to perform statistical analysis of the treatment measures and derive the network relationship and ranking diagrams of the various intervention measures.RESULTS A total of 64 randomized controlled trials involving 5456 patients with UC were included in this study.The adjuvant treatment of UC using five TCM formulations was able to improve the clinical outcome of the patients.Adjuvant treatment with Baitouweng decoction(BTWT)showed a significant effect[mean difference=36.22,95%confidence interval(CI):7.63 to 65.76].For the reduction of tumor necrosis factor in patients with UC,adjunctive therapy with BTWT(mean difference=−9.55,95%CI:−17.89 to−1.41),Shenlingbaizhu powder[SLBZS;odds ratio(OR)=0.19,95%CI:0.08 to 0.39],and Shaoyao decoction(OR=−23.02,95%CI:−33.64 to−13.14)was effective.Shaoyao decoction was more effective than BTWT(OR=0.12,95%CI:0.03 to 0.39),SLBZS(OR=0.19,95%CI:0.08 to 0.39),and Xi Lei powder(OR=0.34,95%CI:0.13 to 0.81)in reducing tumor necrosis factor and the recurrence rate of UC.CONCLUSION TCM combined with mesalazine is more effective than mesalazine alone in the treatment of UC.

Feature extraction for machine learning-based intrusion detection in IoT networks

A large number of network security breaches in IoT networks have demonstrated the unreliability of current Network Intrusion Detection Systems(NIDSs).Consequently,network interruptions and loss of sensitive data have occurred,which led to an active research area for improving NIDS technologies.In an analysis of related works,it was observed that most researchers aim to obtain better classification results by using a set of untried combinations of Feature Reduction(FR)and Machine Learning(ML)techniques on NIDS datasets.However,these datasets are different in feature sets,attack types,and network design.Therefore,this paper aims to discover whether these techniques can be generalised across various datasets.Six ML models are utilised:a Deep Feed Forward(DFF),Convolutional Neural Network(CNN),Recurrent Neural Network(RNN),Decision Tree(DT),Logistic Regression(LR),and Naive Bayes(NB).The accuracy of three Feature Extraction(FE)algorithms is detected;Principal Component Analysis(PCA),Auto-encoder(AE),and Linear Discriminant Analysis(LDA),are evaluated using three benchmark datasets:UNSW-NB15,ToN-IoT and CSE-CIC-IDS2018.Although PCA and AE algorithms have been widely used,the determination of their optimal number of extracted dimensions has been overlooked.The results indicate that no clear FE method or ML model can achieve the best scores for all datasets.The optimal number of extracted dimensions has been identified for each dataset,and LDA degrades the performance of the ML models on two datasets.The variance is used to analyse the extracted dimensions of LDA and PCA.Finally,this paper concludes that the choice of datasets significantly alters the performance of the applied techniques.We believe that a universal(benchmark)feature set is needed to facilitate further advancement and progress of research in this field.